Molecular epidemiology seeks to assess cancer risk in humans by generating hypotheses that can be tested in experimental systems using human tissues. Identification of specific genetic lesions that are important in carcinogenesis is a major effort. In order to adequately address these projects, it is logical to seek consistent genetic mutations in primary tumors and examine their effects in normal or malignant cells in tissue culture. Samples are being collected to generate a primary tumor panel that will be used to determine which gene mutations are worthy of further study or to determine particular spectra of mutations that may relate to specific environmental exposures (e.g., tobacco smoke, industrial effluent, coal smoke or cooking vapors). Expression of specific oncogenes and tumor suppressor genes (erbB-2, p53, ras, transforming growth factor-alpha and epidermal growth factor- receptor) in lung cancer are being studied as potential markers of either early detection or prognosis (survival). Genetic polymorphisms that have been suggested to have an association with human cancer risk are the DNA- restriction fragment length polymorphisms (DNA-RFLPs) at the human HRAS- 1, L-myc, p53, CYP1A1, CYP2E1 and CYP2D6 gene loci. In the case of HRAS- 1, L-myc, CYP1A1, CYP2E1 and p53, rare or minor restriction fragments (alleles) may predispose to certain cancers or may be associated with poor prognosis. For CYP2D6 inheritance rare or mutant forms of the gene may be protective of lung cancer in cigarette smokers. There are no previously existing reports that describe examination of these markers within properly controlled epidemiological studies of lung cancer.